strisunshinewentingwang/pytorch

Pull Requests Commits

Revert D25692616: [pytorch][PR] [reland] Early terminate when CUDA assert were thrown

Rong Rong (AI Infra) committed 4 years ago

69b13735

Merge remote-tracking branch 'origin/master' into ci-all/rohan/test_prof

rohan-varma committed 4 years ago

3ddd827c

Creation of test framework for Sparse Operators (#48488)

hlahkar committed 4 years ago

9552cc65

Revert D25690129: [pytorch][PR] Added linalg.inv

Mike Ruberry committed 4 years ago

5acc27c0

Fix typo in complex autograd docs (#49755)

soulitzer committed 4 years ago

18330092

[reland] Early terminate when CUDA assert were thrown (#49799)

Rong Rong (AI Infra) committed 4 years ago

e6a21559

[numpy] `torch.erfinv`: promote integer inputs to float (#49155)

kshitij12345 committed 4 years ago

3f4b98d5

[pt][quant] Make the CUDA fake quantize logic consistent with CPU fake quantize logic (#49808)

jianyuh committed 4 years ago

4d611093

removes more unused THC functions (#49788)

Natalia Gimelshein committed 4 years ago

e1631729

Improve docs for scatter and gather functions (#49679)

rgommers committed 4 years ago

d99a0c3b

Mod lists to neutral+descriptive terms in caffe2/docs (#49803)

r-barnes committed 4 years ago

b3387139

Added linalg.inv (#48261)

IvanYashchuk committed 4 years ago

8554b58f

Preserve memory format in qconv op (#49533)

Akshit Khurana committed 4 years ago

370350c7

[lint doc] how to fix flake errors if pre-commit hook wasn't there (#49345)

stas00 committed 4 years ago

5171bd94

[Gradient Compression] Directly let world_size = group_to_use.size() (#49715)

Yi Wang committed 4 years ago

55b431b1

[Gradient Compression] Explicitly restrict the scope of torch.cuda.synchronize to the current device (#49711)

Yi Wang committed 4 years ago

88c33ff8

torch.utils.checkpoint.checkpoint + torch.cuda.amp (#49757)

Michael Carilli committed 4 years ago

ee271047

[quant][graphmode][fx] Change standalone module api (#49719)

jerryzh168 committed 4 years ago

f474ffa1

[Gradient Compression] Change wait() to value() in some callbacks of PowerSGD communication hook (#49709)

Yi Wang committed 4 years ago

af1b636b

Add PixelUnshuffle (#49334)

jbschlosser committed 4 years ago

68d438c9

[numpy] `torch.angle`: promote integer inputs to float (#49163)

kshitij12345 committed 4 years ago

461aafe3

Remove unused six code for Python 2/3 compatibility (#48077)

skyline75489 committed 4 years ago

46b83212

Revert D25623219: [pytorch][PR] early terminate when CUDA assert were thrown

Natalia Gimelshein committed 4 years ago

abacf270

Skip None submodule during JIT-tracing (#49765)

Will Feng (DPER) committed 4 years ago

010b9c52

[lint] Apply whitespace linter to all gradle files

Riley Dulin committed 4 years ago

62f9b03b

add type annotations to torch.nn.parallel._functions (#49687)

guilhermeleobas committed 4 years ago

27f0dd36

fx quant: improve types on convert (#49688)

vkuzo committed 4 years ago

de07d076

fx quant: do not observe bias on F.linear (#49628)

vkuzo committed 4 years ago

19f972b6

fx quant: do not observe bias on F.conv (#49623)

vkuzo committed 4 years ago

c3a7591c

Update `is_floating_point()` docs to mention bfloat16 (#49611)

TylerADavis committed 4 years ago

b4141232

Newer Older